Perinatal SSRI Exposure Affects Brain Functional Activity Associated With

www.nature.com/scientificreports

OPEN Perinatal SSRI exposure afects brain functional activity associated with whisker stimulation in adolescent and adult rats Noortje Van der Knaap1, Dirk Wiedermann2, Dirk Schubert1,4, Mathias Hoehn2,4 & Judith R. Homberg1,3,4*

Selective serotonin reuptake inhibitors (SSRI), such as fuoxetine, are used as frst-line antidepressant medication during pregnancy. Since SSRIs cross the placenta the unborn child is exposed to the maternal SSRI medication, resulting in, amongst others, increased risk for autism in ofspring. This likely results from developmental changes in brain function. Studies employing rats lacking the serotonin transporter have shown that elevations in serotonin levels particularly afect the development of the whisker related part of the primary somatosensory (barrel) cortex. Therefore, we hypothesized that serotonin level disturbances during development alter brain activity related to whisker stimulation. We treated female dams with fuoxetine or vehicle from gestational day 11 onwards for 21 days. We investigated ofspring’s brain activity during whisker stimulation using functional magnetic resonance imaging (fMRI) at adolescence and adulthood. Our results indicate that adolescent ofspring displayed increased activity in hippocampal subareas and the mammillary body in the thalamus. Adult ofspring exhibited increased functional activation of areas associated with (higher) sensory processing and memory such as the hippocampus, perirhinal and entorhinal cortex, retrospinal granular cortex, piriform cortex and secondary visual cortex. Our data imply that perinatal SSRI exposure leads to complex alterations in brain networks involved in sensory perception and processing.

According to the World Health Organization, depression is the leading worldwide cause of disability and afects individuals across all life stages, including pregnancy1. Fourteen to sixteen percent of the female population in Te Netherlands reported depressive feelings during their child-bearing years2. Te frst-line treatment for depression involves selective serotonin reuptake inhibitors (SSRI), because of their minimal side-efects3. Of all Dutch pregnant women, 1.8–2% continue antidepressant use during pregnancy and 0.5% start using antidepressants during pregnancy 4. While the SSRI treatment is directed at the mother, by passing the placenta the SSRIs can also reach the unborn child: SSRI levels have been measured in amniotic fuid and breast milk5,6. Tere is evidence that this has consequences for the developing child, as new-borns exhibit withdrawal symptoms afer birth7, called Postnatal Adaption Syndrome, due to a sudden decline in SSRI exposure afer birth. Furthermore, SSRI exposure during pregnancy has been linked to increased risk for pulmonary hypertension, low birth weight, congenital heart defects and defects in motor behaviour 8–12. Additionally, studies focusing on the long-term efects of SSRI exposure on ofspring found associations between prenatal SSRI exposure and increased behavioural problems, such as higher risk for Autism Spectrum Disorder9,13,14. SSRIs block the serotonin transporter (SERT) that re-uptakes serotonin afer it has been released by the syn- apse, lengthening serotonin’s presence in the extracellular space and prolonging its signal 3. Serotonin is also an important regulatory peptide during development with functions in almost the entire brain, including the regulation of synaptogenesis, neuronal outgrowth and migration 15. Terefore, diferent levels of prenatal serotonin can

1Donders Institute for Brain, Cognition and Behaviour, Radboud University and Radboud University Medical Center, 6500 HB Nijmegen, The Netherlands. 2In‑Vivo‑NMR Laboratory, Max Planck Institute for Metabolism Research, Cologne, Germany. 3Department of Cognitive Neuroscience, Donders Institute for Brain, Cognition and Behaviour, Radboud University Nijmegen Medical Centre, Kapittelweg 29, 6525 EN Nijmegen, The Netherlands. 4These authors jointly supervised this work: Dirk Schubert, Mathias Hoehn and Judith R. Homberg. *email: [email protected]

Scientifc Reports | (2021) 11:1680 | https://doi.org/10.1038/s41598-021-81327-z 1 Vol.:(0123456789) www.nature.com/scientificreports/

Figure 1. Model of the experimental timeline and set-up. GD gestational days, PND postnatal days.

afect the brains’ development. Long-term controlled studies in humans are challenging because of the ethical difculties of randomized controlled studies. Rodent models fll this gap, giving us the opportunity to closely investigate the consequences of developmental SSRI exposure. Tese preclinical studies have suggested diferences in social behaviour, learning and memory, anxiety and sensory functions, alterations in neuronal organization, diferences in hippocampal neurogenesis and epigenetic changes afer SSRI exposure during pregnancy and the early postnatal phase16–28. Several studies in rodents have demonstrated that elevated serotonin levels during brain development lead to prominent alterations in the organisation of the somatosensory system, in particular in the thalamocortical innervation and intracortical networks of the primary somatosensory (barrel) cortex. Perinatal SSRI exposure has been associated with reduced interhemispheric barrel cortex connectivity29. In rodents lacking SERT (SERT knockout) the barrel cortex is characterized by changes in the size and width of the barrels, as well as reduced inhibitory control over excitatory neurons and thereby probably altered sensory gating30–33. Furthermore, SERT knockout rats were found to be faster in sensory integration in the gap crossing task in which the animals use their whiskers to cross a gap between two platforms30. Further, cerebral glucose utilization was found to be reduced in the whisker-to-barrel cortex pathway of SERT knockout mice compared to control mice upon unilateral whisker stimulation 34. Since there are similarities in the developmental consequences of SERT knockout and perinatal SSRI exposure35,36, and perinatal SSRI exposure afects sensory functions as well as behaviours dependent on sensory processing (see above), we hypothesize that perinatal SSRI exposure afects the activity of brain areas associated with the somatosensory system later in life. To test this hypothesis, we sought to investigate the efect of perinatal SSRI (fuoxetine) exposure on the later life brain-wide response to whisker stimulation, using functional magnetic resonance imaging (fMRI). We have chosen this approach to maximize the translational fndings to humans. More specifcally, we treated dams from gestational day 11 until postnatal day 11 with fuoxetine. We have chosen this period as serotonergic neurons emerge at embryonic day 10–12 in rodents 37, and postnatal day 2 to 11 constitutes a 5-HT-sensitive period impacting (sensory-dependent) anxiety-like behaviour in ofspring 38. Additionally, as postnatal day 10–11 in the rat corresponds to birth in humans39, with this treatment period we model SSRI exposure approximately during the second and third trimester of human pregnancy. In the ofspring we studied how this perinatal SSRI exposure afected brain activity patterns associated with whisker stimulation at adolescence and adulthood. Material and methods Animals. Twelve female and fve male Wistar rats (Harlan-Winkelmann GmbH, Borchen, Germany, and Janvier, Le Genest Saint Isle, Cedex, France) were housed under controlled temperature conditions (21 ± 1 °C), humidity (55% ± 10%) and light/dark cycle (12/12 h) at the Max Planck Institute of Metabolism Research, Cologne. Tey were given access to food and water ad libitum. Afer 2 weeks of acclimation female rats were measured daily for their oestrous stage with an impedance checker (Impedance Checker MK-10B, Muromachi Kikai, Tokyo, Japan). When the females were in oestrous, they were paired with a male and placed in a cage with a custom-made wire bottom (stainless steel, containing small circular holes of 10 mm, spacing 13 mm). Twelve hours later male and female were separated and the cage was inspected for the vaginal plug. If found, the day was marked as gestational day (GD) 1. Te female dams were handled daily and treated during the exposure period, via an oral gavage from GD 11 onwards for 21 days with 12 mg/kg fuoxetine or 1% methylcellulose (vehicle). In total 6 dams were treated with fuoxetine and 6 dams with vehicle. Pups derived from these dams were weaned at 3 weeks and inspected for sex. Male pups proceeded to the MRI procedures which were in total 26 animals for the control group and 15 for the SSRI group. All animals underwent two MRI periods (functional and structural) on two days at 5 and 10 weeks, representing adolescence and adulthood respectively (Fig. 1). Due to the long acquisition times and the need of diferent anaesthetics, the functional and the structural scans could not be performed in the same scanning session. Consequently, on day one a dedicated fMRI session was conducted, and on day two an anatomical scan was made. Tree animals of the fuoxetine group and one animal in the control group were lost during MRI scans. All animal experiments were carried out in accordance with the guidelines of the German Animal Welfare Act and approved by the ethical committee for animal studies of the local authori- ties (Landesamt für Natur, Umwelt und Verbraucherschutz Nordrhein-Westfalen).

Scientifc Reports | (2021) 11:1680 | https://doi.org/10.1038/s41598-021-81327-z 2 Vol:.(1234567890) www.nature.com/scientificreports/

Magnetic resonance imaging. MRI experiments were conducted on a 9.4 T horizontal bore dedicated animal scanner (Biospec 94/20, Bruker Biospin, Ettlingen, Germany). Radiofrequency transmission was achieved with a quadrature volume resonator (inner diameter 72 mm), and a rat brain surface coil (~ 30 × 30 mm2) (Bruker, Biospin) was used for signal detection. A custom-made cradle was used (Medres, Cologne, Ger- many) to perform the fMRI whisker stimulation. Te head was fxated using a bite bar and ear bars to reduce movement artefacts. Respiration, heart rate and body temperature were controlled at all times during scanning. Body temperature was maintained at 37 °C via a controlled water circulation system integrated in the cradle (Medres, Cologne, Germany). Imaging was conducted with the Paravision sofware, version 5.1 (Bruker Bio- Spin). Te animals were carefully positioned to make sure that the barrel cortex was aligned with the isocenter of the magnet. To optimize feld homogeneity, the implemented MAPSHIM of Paravision was used for shimming locally for the barrel cortex. For fMRI a gradient echo EPI sequence was used. TR/TEef = 3,502 ms/17.5 ms; feld of view 28.8 mm × 28.8 mm (matrix 96 × 96) and 1.2 mm slice thickness; pixel size = 0.3 × 0.3 mm; 11 slices. Te T2-weighted imaging (T2WI) was acquired using a Multi Slice Multi (Spin-) Echo sequence (MSME; TR/TE = 5500 ms/10 ms; 16 echoes, slice thickness = 0.5 mm; 30 slices; in-plane resolution = 0.145 × 0.145 mm2. For the generation for anatomical images for registration within the fMRI procedure we used the TurboRARE sequence with a TR/TEef = 6500 ms/32.5 ms; feld of view 32 mm × 32 mm (matrix 256 × 256) and 0.5 mm slice thickness; voxel size = 0.3 × 0.3 mm2; 11 slices.

MRI scanning procedure. Te MRI protocol contained a 2-day procedure and was conducted at 5 and 10 weeks of age. On day one the fMRI protocol was run. Te animals were anesthetized with 2% isofurane (Forane, Baxter, Deerfeld, IL, USA) in a 70/30 mixture of N20 and O 2 and placed within the MRI cradle. Te whiskers were prepped for fMRI whisker stimulation where, unilaterally, the whisker B1, C1, D1 and E1 were fagged together with tape while the other whiskers were carefully fxated away from the airfow. A tube with N2 was fxed to hit the fag in the middle and was tested with each animal for correct positioning before placing the animals within the scanner. Te N2O was then replaced by N2. Te animal was given a 0.5 ml bolus of 0.05 mg/kg of medetomidine (Domitor, Pfzer, Karlsruhe, Germany; 1 ml/kg in 10 ml of saline) subcutaneously40. Isofurane was slowly discontinued and sedation was maintained via continuous subcutaneous infusion of a 1 ml/h metho- midine solution (0.1 mg/kg). Te fMRI procedure was conducted approximately 30 min afer bolus injection. Te TurboRARE was recorded for anatomical realignment of the fMRI image before the fMRI procedure was executed. Te whiskers were stimulated with a N2 puf in an anterior to posterior defection with a frequency of 6 Hz and a block design consisting of 5 blocks of activation of 15 s each intertwined with 45 s of no stimulation. Te long periods of no stimulation were included to make sure that the somatosensory system would not desen- sitize due to overstimulation. Afer the imaging session, atipamezol (Anitsedan, Pfzer, 1 ml/kg bodyweight) was injected subcutaneously, together with 2 ml of saline, to reverse the sedative efect and overcome any fuid loss due to the diuretic efect of medetomidine solution. Twenty-four hours later animals were placed in the scanner again, this time under isofurane anaesthetic without any whisker preparation for structural T2 weighted scans.

Functional MRI analysis. Successful scans for the fMRI analysis were obtained from 14 control animals and 12 perinatally SSRI exposed animals at 5 weeks of age, and 18 control animals and 11 perinatally SSRI exposed animals at 10 weeks of age. Tese animals were used for further image analyses. Te fMRI images were converted to NIfTI format and voxel size in the header was scaled with a factor 10 to account for the brain size diferences between rat and humans 41. FSL version 4.0.542 was used for all analysis unless stated otherwise. Images were pre-processed, with rigid body motion correction and temporal high pass fltering. Recorded breathing, cardiac and movement information obtained was used to regress out non-specifc Blood Oxygen Level Dependent (BOLD) activation. fMRI images were aligned to its own structural image. Te structural images were used to align to the in-house rat template image by means of ANTS 43,44. Statistical analysis where performed in two steps with FEAT (fMRI expert analysis tool of FSL version 6.0). At the frst level statistics, the statistical activation maps were calculated using general linear model (GLM) for the block design of whisker stimulation using the convolution of double-gamma Hemodynamic Response Function. GLM uses FILM (FMRIB’s improved linear model) to pre-whiten each voxel time series, to create robust estimation time-series (described in more detail elsewhere 45). To compare the SSRI and vehicle treated groups at 5 and 10 weeks, we preformed FEAT’s second level comparative statistic with FLAME (FMRIB’s Local Analysis of Mixed Efects) within FEAT, upon the statistical actions maps from frst level statistics to compare diferences in activation patterns. FLAME used Marko Chain Monte Carlo sampling to get an accurate estimation of the true random-efects variance and degrees of freedom at each voxel (as described in more detail elsewhere 46). Upon the z-level statistical images, a mask of the whole brain in-house rat template (as described above) was used to exclude any voxels outside of the brain in the comparison, and a cluster level corrected statistic was performed with a z threshold of 2.4 and a cluster P threshold of < 0.0547.

Voxel based morphometry analysis. Successful scans for the T2 weighted scans were obtained at 5 weeks of age from 15 control and 10 perinatally SSRI exposed animals, and at 10 weeks of age from 17 control and 11 perinatally SSRI exposed animals. Like for the fMRI images, the T2 weighted images were converted to NIFTI format and resized with a factor of 10 to account the diferences between human and rodent brain size. Te procedures were further executed with SPM 12 (Welcome Trust Centre for Neuroimaging, London, UK). Te mean of the T2 weighted images were realigned and then rigid body aligned to the high resolution Wistar atlas image48 and resized to 1.25 mm isotropic voxel. Te image was then segmented (warping regularization: 1; warp frequency cut-of: 25; bias regularization:

Scientifc Reports | (2021) 11:1680 | https://doi.org/10.1038/s41598-021-81327-z 3 Vol.:(0123456789) www.nature.com/scientificreports/

Figure 2. Te efect of whisker stimulation block > control block over all subject groups. In both the perinatally SSRI exposed and control groups whisker stimulation to the lef side is associated with increased activation of the barrel cortex at the contralateral side (5 weeks: z-max = 6.43, p < 0.001; 10 weeks: z-max is 6.25; p < 0.001). Additionally, there is strong activation in thalamic projection sides and sensory relay regions like the ventral posteromedial nucleus and posteromedial complex region. Cluster activation is shown on an in-house rat template and z-scores are depicted on the scale bar.

0.0001; bias FWHM: 60 mm cut-of; and sampling distance: 3)) to GM (gray matter), WM (white matter) and CSF (cerebrospinal fuid) using the rat atlas tissue priors 48. To obtain the most accurate registration of the segmented images, DARTEL (Difeomorphic Anatomical Registration Trough Exponentiated Lie algebra) was used for an automated, unbiased, and nonlinear template building 49. Afer the DARTEL template building procedure the images were normalized either non-modulated or modulated to its own DARTEL age specifc template and smoothed with an 8 × 8 × 8 mm FWHM Gaussian kernel. Both modulation approaches were used to inspect for volume and density diferences between groups. A general linear model was conducted to compare the diferences of treatment for each age group. We corrected for whole brain volume as global measurement. Te images were thresholded with an exact mask of 0.2 to investigate the contrasts of control > perinatal SSRI exposure and perinatal SSRI > control exposure with a family wise error correction of p < 0.05. Results Functional MRI. Generally, unilateral stimulation of one row of whiskers led to a most prominent and strong activation in the contralateral barrel feld of the rat (Fig. 2). Te stimulus induced activation pattern was revealed by means of the overall contrast of the whisker stimulation block over the control block and at both 5 (n = 26) and 10 (n = 29) weeks of age we observed large signifcant clusters of activation afer correction for multiple comparisons (5 weeks: z-max = 6.43, p < 0.001; 10 weeks: z-max is 6.25; p < 0.001). However, fMRI also showed that the stimulation and resulting movement of the whiskers not only activated the barrel cortex. In 5-week-old rats, we found the peak maxima of the activity also in somatosensory nuclei of the thalamus, i.e. the thalamic ventral posteromedial nucleus (VPM) and posterior Talamic (Po) group 50,51. Furthermore, the stimulation procedure resulted in the activation of non-somatosensory areas, in particular in the primary auditory area. At 10 weeks of age the peak maxima were located in similar areas as at 5 weeks of age, including the barrel cortex, the VPM and Po region of the thalamus and in the primary auditory area. However, unlike at the younger age, whisker stimulation also led to peak maxima in medial secondary visual cortex, parietal association area, and perirhinal and entorhinal cortex (Table 1). When comparing whisker stimulation evoked activity in brains of perinatal SSRI exposed animals at 5 weeks of age over the respective controls (perinatal SSRI exposure > control), we found no signifcant diferences in barrel feld activation. However, we did fnd diferences in the CA3 and dentate gyrus subareas of the hippocampus, stretching to the mammillary body in the thalamus (cluster size = 72 voxels, z-max 1.9, p = 0.012) (Fig. 3). We found no diferences for the contrast control > perinatal SSRI exposure. Diferences in barrel feld activation were also not detectable between SSRI exposed brains and controls at 10 weeks of age. As in the 5-week-old animals, we did observe an increased signal in the perinatally SSRI exposed group over the control group in several other brain regions. However, the regions were to a large extent diferent from those found at 5 weeks of age (Fig. 4). Six clusters of activation survived whole-brain cluster- correction. As a whole, these detected regions are known to be involved in sensory information processing, sensory integrating, and memory. Most anterior we found increased activation of the piriform cortex (cluster size = 54 voxels z-max = 3.81, p = 0.003). Posterior to this and overlapping in location we observed activation in an area that was spanning from motor cortex (M1) to the primary somatosensory cortex (S1) (cluster size = 38 voxels, z-max = 3.0, p = 0.02). Around the Bregma location of AP − 2.5 we observed activation in the retrospinal (a)granular cortex (cluster size = 38 voxels, z-max = 3.2, p = 0.03). Most posterior (Bregma AP = − 5.0) we found three clusters. One cluster in the lef CA2 hippocampal region (cluster size = 87 voxels, z-max = 4.93, p < 0.001), one cluster that is located in the perirhinal cortex and entorhinal region stretching into the right hippocampal

Scientifc Reports | (2021) 11:1680 | https://doi.org/10.1038/s41598-021-81327-z 4 Vol:.(1234567890) www.nature.com/scientificreports/

z-value Area 6.43 Barrel cortex 6.23 Barrel cortex 5.96 Barrel Cortex 5 weeks 5.91 Barrel cortex 5.91 Ventral posteromedial thalamic nucleus and Posterior Talamic group 5.86 Secondary auditory cortex 6.25 Medial secondary visual cortex 6.12 Associative parietal cortex 6.11 Barrel cortex 10 weeks 6.10 Primary auditory cortex, ectorhinal and perirhinal cortex 5.84 Primary Auditory Cortex, Ectorhinal And Perirhinal cortex 5.84 Ventral posteromedial thalamic nucleus and Posterior Talamic group

Table 1. Peak Maxima of overall contrast whiskers vs no whiskers stimulation p < 0.001.

Figure 3. Results of the contrast perinatal SSRI exposure > control for the 5 weeks old rats. Whisker stimulation was given on the lef side of the animal. Increased activity is observed in the hippocampal CA3 and dentate gyrus regions stretching to the mammillary body in the thalamus (cluster size = 72 voxels, z-max 1.9 p = 0.012). Signifcant clusters are shown upon an in-house rat template and z-scores are depicted in the scale-bar.

CA2 region (cluster size = 60 voxels, z-max = 4.07, p = 0.002), and lastly an activation cluster in the secondary visual cortex (cluster size = 35 voxels, z-max = 3.63 p = 0.04). Tere is no cluster signifcant in the contrast control > perinatal SSRI exposure.

Voxel based morphometry (VBM). At 5 and 10 weeks of age no signifcant efects (p > 0.05) were found between the SSRI treated group and the control group for either volume or density diferences in the voxel-based morphometry analysis. Discussion In this study we set out to test how perinatal exposure to the SSRI fuoxetine afects large-scale functional brain activity associated with tactile sensory activation. We measured in adolescent and adult rats functional brain activity using fMRI during whisker stimulation, to obtain insight in the later life consequences of perinatal SSRI exposure. Rodents show large scale activation of the barrel cortex upon whisker stimulation, but no diferences in the BOLD response in the barrel feld were found between the control and perinatally SSRI exposed groups. However, in several regions involved in higher-level processing of sensory information or memory integration we observed increased activation in the perinatally SSRI exposed group compared to the control group, particularly in adult rats. Te long-term efects of prenatal SSRI exposure on human ofspring have been shown to involve increased behavioural problems in later life, including a higher risk for Autism Spectrum Disorders 9,13,14, which are characterized by altered sensory integration and function 52–54. However, this association was not always replicated 55–57. It has been reported that perinatal SSRI exposure in humans afects the function of the auditory cortex. Tese fndings may relate to serotonin’s role in the regulation of sensory processes 58,59, and the association between serotonergic genetic mutations and autism-related phenotypes as observed in animals 60. Tis link of serotonin with sensory processing alterations and outcome is supported by our fndings that early alterations in serotonin

Scientifc Reports | (2021) 11:1680 | https://doi.org/10.1038/s41598-021-81327-z 5 Vol.:(0123456789) www.nature.com/scientificreports/

Figure 4. Results of the activated brain clusters in the 10-week old rats. Te contrast for perinatal SSRI exposure over the control condition is shown. Whisker stimulation was given on the lef side of the animal. Six diferent signifcant clusters are shown. Te yellow square shows the contralateral activation of the piriform cortex (cluster size = 54 voxels, z-max = 3.81, p = 0.003). In both the yellow and red squares, the activation of the cluster encompassing the motor cortex and primary sensory cortex is shown (cluster size = 38 voxels, z-max = 3.0, p = 0.02). In the blue square the activation of the cluster in the retrosplenal granular and agranular cortex is shown (cluster size = 38 voxels, z-max = 3.2, p = 0.03). In the green square the activation of three clusters is shown: Ipsilateral to the stimulation side and most dorsal the secondary visual cortex (cluster size = 35 voxels, z-max = 3.63 p = 0.04), and the CA2 of the hippocampus (cluster size = 87 voxels, z-max = 4.93 and p < 0.001). Lastly, a cluster covering the perirhinal and entorhinal cortex stretching into the hippocampal CA2 region is activated (cluster size = 60 voxels, z-max = 4.07, p = 0.002). All clusters shown survived whole brain cluster correction and are laid over an in-house rat template. Value-bar depicts the z-value range.

signalling due to SSRI exposure afect the integration of sensory information. Tis integration could on the long-term afect behaviour related to the autistic spectrum. When inspecting our results, we did not observe in the barrel cortex itself signifcant diferences in overall activation between exposure groups, although we specifcally evoked whisker-related sensory input. It is possible that the limited spatial and time resolution of fMRI rendered small changes in the barrel cortex undetect- able. It is also possible that over time, during development, changes in barrel cortex function due to perinatal SSRI exposure dissipated or restored but led to down-stream changes in other brain regions. Te capability of the sensory cortical areas to compensate for structural distortions was shown in mouse models with severely distorted laminar organisation of the cortex61. Even in the absence of any normal structural correlate to barrels, imaging of whisker stimulation related activity revealed reinstated topological network recruitment 62. A developmental normalization of barrel cortex function could explain why the general BOLD response in the barrel cortex is similar between groups at adolescent and adult age. Potentially, the plastic brain rewires when the serotonin levels are gradually normalized afer postnatal day 11, when SSRI exposure has stopped. However, SSRI exposure still results in diferential functional activation associated with sensory stimulation, but refected in other brain regions. When comparing the activation of whisker stimulation as main efect, adolescent and adult rodents showed a diferent pattern of peak maxima. Te peak maxima in the adolescent group are mainly located in the barrel cortex, while the adults show activation in other sensory regions, like visual and auditory areas (Table 1)63. In the diferent exposure groups, perinatal SSRI over control, the adolescent group displayed activity diferences in the hippocampus and mammillary body. In the adult rats, perinatal fuoxetine exposure led to more BOLD activity compared to vehicle exposure in response to whisker stimulation in other sensory regions than the barrel cortex. We observed activation in the piriform cortex which is involved in olfaction. In addition, we observed activation in the motor cortex and primary somatosensory cortex, secondary visual cortex and hippocampus. Moreover, we detected increased activation in the retrospinal granular cortex, involved in spatial judgement, the perirhinal cortex involved in sensory information processing, and the enthorinal cortex, involved in spatial mapping (Fig. 4). All these areas are implicated in the integration of environmental sensory information. Tese results were not due to diferences in brain size, as shown by the VBM analysis. Te diferent activation patterns at adolescence and adulthood for the main task efect may be indicative of more integration of sensory input across development. Multisensory integration develops over age in humans, mostly before adulthood64–66. Multi-sensory integration is more pronounced at adult ages than during childhood, and used more ofen by adults than by children 67. Our fndings support this developmental view. Our results are also in line with studies in which tactile information was disrupted by removing whiskers during neonatal period and multi-sensory integration diferences were found well into adulthood68. Tis pattern of results indi- cates that sensory (integration) systems are highly dynamic during development, able to overcome distortions

Scientifc Reports | (2021) 11:1680 | https://doi.org/10.1038/s41598-021-81327-z 6 Vol:.(1234567890) www.nature.com/scientificreports/

to a certain extent. However, they always have to recalibrate within the limits of wiring and experience. Whether the increased activity in the diferent sensory processing regions as we found in the present study is a sign of a still developing sensory integration process or an indication that the brain is diferentially calibrated due to the developmental setback afer fuoxetine exposure, is unknown. Nonetheless, the present data provide initial insight into the brain-wide efects of the maternal use of SSRI on sensory processing. Previous research revealed that rodents prenatally exposed to an SSRI exhibited decreased spatial memory performance 22. Our results suggest that mostly hippocampal and memory-related regions show increased activation in perinatally fuoxetine exposed ofspring compared to controls. Tis could be due to an implicit increase in memory integration or a result of increased activity in diferent sensory regions important for the integration of information and subsequent memory encoding. Tis work has some limitations. Due to strict regulations in animal testing and the explorative nature of this experiment, we investigated a restricted number of animals in this study. Nonetheless, the present study shows that fMRI is a valid tool to investigate large-scale brain functional and structural changes due to developmental perturbations in serotonin levels. Another limitation is that the female rats treated with SSRIs were healthy, while in humans SSRIs are only provided to women having depressive symptoms. We therefore cannot draw conclu- sions regarding the potential adverse efects of maternal depression on brain-wide activity patterns in response to sensory stimulation. Further, while the dams were handled very gently before treatment, we cannot rule out the possibility that the oral gavage stressed the dams, which in turn could infuence ofspring brain development. Also, as we only tested the SSRI fuoxetine, it remains unknown whether fndings also generalize to other SSRIs. Finally, we cannot exclude the possibility that the blue square in Fig. 4 presents the venus sagitalis rather than the retrosplenal and agranular cortex, as the fMRI signal in this position has previously been suggested to stem from this vene69. We would also like to point out that the anaesthesia we needed to apply to the animals during the imaging may have had efects on the stimulus evoked brain activity. However, based on previous reports70 we can expect that the resulting brain activity patterns closely refected those of awake animals and that the anaesthesia was not a major confounding factor for detecting diferences in the responses of untreated and SSRI exposed ofspring. We never experience our environment with one sensory system only. Te integration of visual, tactile, ves- tibular, proprioceptive, auditory and olfactory information is how we structure the world and our place in it. Te integration of sensory input and the integration of this input with memory processes is key to our ftness and known to develop well into adulthood 64. Exposure to serotonin altering medication during development has been related to diferences in neuronal and behavioural levels 16–28. In this study we found brain-wide diferences in functional activity during adolescence and adulthood as a consequence of perinatal SSRI exposure. Our fndings suggest that developmental SSRI exposure afects sensory processing of the brain even into adulthood. Hence, future research should not focus only on primary sensory areas, but also examine the overarching picture of sensory integration in the entire brain afer developmental disrupting infuences, like antidepressant medication.

Received: 14 February 2020; Accepted: 5 January 2021

References 1. Organization, G. W. H. Depression and other common mental disorders: Global health estimates. License: CC BY-NC-SA 3.0 IGO (2017). 2. Verweij, G. & Houben-van Herten, M. Depressiviteit en antidepressiva in Nederland. Centraal Bureau voor de Statistiek (2013). 3. Frazer, A. Antidepressants: Pharmacology and mechanisms of action. J. Clin. Psychopharmacol. 17, 1–33 (1997). 4. Ververs, T. et al. Prevalence and patterns of antidepressant drug use during pregnancy. Eur. J. Clin. Pharmacol. 62, 863–870 (2006). 5. Hendrick, V. et al. Placental passage of antidepressant medications. Am. J. Psychiatry 160, 993–996 (2003). 6. Kim, J. et al. Stereoselective disposition of fuoxetine and norfuoxetine during pregnancy and breast-feeding. Br. J. Clin. Pharmacol. 61, 155–163 (2006). 7. Sanz, E. J., De-las-Cuevas, C., Kiuru, A., Bate, A. & Edwards, R. Selective serotonin reuptake inhibitors in pregnant women and neonatal withdrawal syndrome: A database analysis. Lancet (London, England) 365, 482–487 (2005). 8. Gentile, S. & Galbally, M. Prenatal exposure to antidepressant medications and neurodevelopmental outcomes: A systematic review. J. Afect. Disord. 128, 1–9 (2011). 9. Hanley, G. E., Brain, U. & Oberlander, T. F. Prenatal exposure to serotonin reuptake inhibitor antidepressants and childhood behavior. Pediatr. Res. 78, 174–180 (2015). 10. Noorlander, C. W. et al. Modulation of serotonin transporter function during fetal development causes dilated heart cardiomyo- pathy and lifelong behavioral abnormalities. PLoS ONE 3, e2782 (2008). 11. Pedersen, L. H., Henriksen, T. B. & Olsen, J. Fetal exposure to antidepressants and normal milestone development at 6 and 19 months of age. Pediatrics 125, e600–e608 (2010). 12. Suri, R., Lin, A. S., Cohen, L. S. & Altshuler, L. L. Acute and long-term behavioral outcome of infants and children exposed in utero to either maternal depression or antidepressants: A review of the literature. J. Clin. Psychiatry https://doi.org/10.4088/JCP.13r08 926 (2014). 13. El Marroun, H. et al. Prenatal exposure to selective serotonin reuptake inhibitors and social responsiveness symptoms of autism: Population-based study of young children. Br. J. Psychiatry 205, 2 (2014). 14. Croen, L. A., Grether, J. K., Yoshida, C. K., Odouli, R. & Hendrick, V. Antidepressant use during pregnancy and childhood autism spectrum disorders. Arch. Gen. Psychiatry https://doi.org/10.1001/archgenpsychiatry.2011.73 (2011). 15. Chameau, P. et al. Te N-terminal region of reelin regulates postnatal dendritic maturation of cortical pyramidal neurons. Proc. Natl. Acad. Sci. U. S. A. 106, 7227–7232 (2009). 16. Boulle, F. et al. Prenatal stress and early-life exposure to fuoxetine have enduring efects on anxiety and hippocampal BDNF gene expression in adult male ofspring. Dev. Psychobiol. 58, 427–438 (2016). 17. Gemmel, M. et al. Perinatal fuoxetine has enduring sexually diferentiated efects on neurobehavioral outcomes related to social behaviors. Neuropharmacology 144, 70–81 (2019). 18. Gemmel, M. et al. Gestational stress and fuoxetine treatment diferentially afect plasticity, methylation and serotonin levels in the PFC and hippocampus of rat dams. Neuroscience 327, 32–43 (2016).

Scientifc Reports | (2021) 11:1680 | https://doi.org/10.1038/s41598-021-81327-z 7 Vol.:(0123456789) www.nature.com/scientificreports/

19. Gemmel, M. et al. Perinatal fuoxetine increases hippocampal neurogenesis and reverses the lasting efects of pre-gestational stress on serum corticosterone, but not on maternal behavior, in the rat dam. Behav. Brain Res. 339, 222–231 (2018). 20. Gemmel, M. et al. Perinatal fuoxetine efects on social play, the HPA system, and hippocampal plasticity in pre-adolescent male and female rats: Interactions with pre-gestational maternal stress. Psychoneuroendocrinology 84, 159–171 (2017). 21. Gingrich, J. A. et al. New insights into how serotonin selective reuptake inhibitors shape the developing brain. Birth Defects Res. 109, 924–932 (2017). 22. Sprowles, J. L. N. et al. Diferential efects of perinatal exposure to antidepressants on learning and memory, acoustic startle, anxiety, and open-feld activity in Sprague-Dawley rats. Int. J. Dev. Neurosci. 61, 92–111 (2017). 23. Maloney, S. E. et al. Examining the reversibility of long-term behavioral disruptions in progeny of maternal ssri exposure. eNeuro 5, 1–27 (2018). 24. Zahra, A., Jiang, J., Chen, Y., Long, C. & Yang, L. Memantine rescues prenatal citalopram exposure-induced striatal and social abnormalities in mice. Exp. Neurol. 307, 145–154 (2018). 25. Houwing, D. J., Heijkoop, R., Olivier, J. D. A. & Snoeren, E. M. S. Perinatal fuoxetine exposure changes social and stress-coping behavior in adult rats housed in a seminatural environment. Neuropharmacology 151, 84–97 (2019). 26. Silva, A. S. et al. Maternal exposure to fuoxetine during gestation and lactation induces long lasting changes in the DNA methylation profle of ofspring’s brain and afects the social interaction of rat. Brain Res. Bull. 142, 409–413 (2018). 27. Bond, C. M. et al. Perinatal fuoxetine exposure results in social defcits and reduced monoamine oxidase gene expression in mice. Brain Res. 1727, 146282 (2020). 28. Yu, W. et al. Prenatal selective serotonin reuptake inhibitor (SSRI) exposure induces working memory and social recognition defcits by disrupting inhibitory synaptic networks in male mice. Mol. Brain 12, 1–13 (2019). 29. Simpson, K. L. et al. Perinatal antidepressant exposure alters cortical network function in rodents. PNAS https://doi.org/10.1073/ pnas.1109353108 (2011). 30. Miceli, S. et al. Reduced inhibition within layer IV of sert knockout rat barrel cortex is associated with faster sensory integration. Cereb. Cortex 27, 933–949 (2017). 31. Miceli, S. et al. High serotonin levels during brain development alter the structural input–output connectivity of neural networks in the rat somatosensory layer IV. Front. Cell. Neurosci. 7, 88–103 (2013). 32. Akhmetshina, D. et al. Te serotonin reuptake inhibitor citalopram suppresses activity in the neonatal rat barrel cortex in vivo. Brain Res. Bull. 124, 48–54 (2016). 33. Chen, X. et al. Disruption of transient serotonin accumulation by non-serotonin-producing neurons impairs cortical map development. Cell Rep. 10, 346–358 (2015). 34. Esaki, T. et al. Developmental disruption of serotonin transporter function impairs cerebral responses to whisker stimulation in mice. Proc. Natl. Acad. Sci. U.S.A. 102, 5582–5587 (2005). 35. Homberg, J. R., Schubert, D. & Gaspar, P. New perspectives on the neurodevelopmental efects of SSRIs. Trends Pharmacol. Sci. 31, 60–65 (2009). 36. Kroeze, Y. et al. Perinatal reduction of functional serotonin transporters results in developmental delay. Neuropharmacology 109, 96–111 (2016). 37. Gaspar, P., Cases, O. & Maroteaux, L. Te developmental role of serotonin: News from mouse molecular genetics. Nat. Rev. Neurosci. 4, 1002–1012 (2003). 38. Rebello, T. J. et al. Postnatal day 2 to 11 constitutes a 5-HT-sensitive period impacting adult mPFC function. J. Neurosci. 34, 12379–12393 (2014). 39. Workman, A. D., Charvet, C. J., Clancy, B., Darlington, R. B. & Finlay, B. L. Modeling transformations of neurodevelopmental sequences across mammalian species. J. Neurosci. 33, 7368–7383 (2013). 40. Weber, R., Ramos-Cabrer, P., Wiedermann, D., Van Camp, N. & Hoehn, M. A fully noninvasive and robust experimental protocol for longitudinal fMRI studies in the rat. Neuroimage 29, 1303–1310 (2006). 41. Bajic, D., Craig, M. M., Mongerson, C. R. L., Borsook, D. & Becerra, L. Identifying rodent resting-state brain networks with inde- pendent component analysis. Front. Neurosci. 11, 2 (2017). 42. Jenkinson, M., Beckmann, C. F., Behrens, T. E. J., Woolrich, M. W. & Smith, S. M. Fsl. Neuroimage 62, 782–790 (2012). 43. Avants, B. B., Tustison, N. J., Wu, J., Cook, P. A. & Gee, J. C. An open source multivariate framework for n-tissue segmentation with evaluation on public data. Neuroinformatics 9, 381–400 (2011). 44. Avants, B. B. et al. A reproducible evaluation of ANTs similarity metric performance in brain image registration. Neuroimage 54, 2033–2044 (2011). 45. Woolrich, M. W., Ripley, B. D., Brady, M. & Smith, S. M. Temporal autocorrelation in univariate linear modeling of FMRI data. Neuroimage 14, 1370–1386 (2001). 46. Woolrich, M. W., Behrens, T. E. J., Beckmann, C. F., Jenkinson, M. & Smith, S. M. Multilevel linear modelling for FMRI group analysis using Bayesian inference. Neuroimage 21, 1732–1747 (2004). 47. Worsley, K. J. in Functional MRI: An Introduction to Methods. in Statistical analysis of activation images (eds. Jezzard, P., Matthews P.M. & Smith, S. M.) (OUP, 2001). 48. Valdés-Hernández, P. A. An in vivo MRI template set for morphometry, tissue segemantaion, and fMRI localization in rats. Front. Neuroinform. 5, 1–19 (2011). 49. Ashburner, J. A fast difeomorphic image registration algorithm. Neuroimage 38, 95–113 (2007). 50. Bosman, L. W. J. et al. Anatomical pathways involved in generating and sensing rhythmic whisker movements. Front. Integr. Neurosci. 5, 1–28 (2011). 51. Kleinfeld, D. & Deschenes, M. Neuronal basis for object location in the vibrissa scanning sensorimotor system. Neuron 72, 455–468 (2011). 52. Tye, M. D., Bednarz, H. M., Herringshaw, A. J., Sartin, E. B. & Kana, R. K. Te impact of atypical sensory processing on social impairments in autism spectrum disorder. Dev. Cogn. Neurosci. 29, 151–167 (2018). 53. Zhou, H. et al. Multisensory temporal binding window in autism spectrum disorders and schizophrenia spectrum disorders: A systematic review and meta-analysis. Neurosci. Biobehav. Rev. https://doi.org/10.1016/j.neubiorev.2017.12.013 (2018). 54. Bedford, R., Pellicano, E., Mareschal, D. & Nardini, M. fexible integration of visual cues in adolescents with autism spectrum disorder. Autism Res https://doi.org/10.1002/aur.1509 (2016). 55. Hviid, A. & Melbye, M. Use of selective serotonin reuptake inhibitors during pregnancy and risk of autism. N. Engl. J. Med. https ://doi.org/10.1056/NEJMoa1301449 (2013). 56. Sørensen, M. J. et al. Antidepressant exposure in pregnancy and risk of autism spectrum disorders. Clin. Epidemiol. 5, 449–459 (2013). 57. Zhou, X., Li, Y., Ou, J. & Li, Y. Association between maternal antidepressant use during pregnancy and autism spectrum disorder: An updated meta-analysis. Mol. Autism. 9, 1–7 (2018). 58. Deemyad, T., Metzen, M. G., Pan, Y. & Chacron, M. J. Serotonin selectively enhances perception and sensory neural responses to stimuli generated by same-sex conspecifcs. Proc. Natl. Acad. Sci. 110, 19609–19614 (2013). 59. Lottem, E., Lorincz, M. L. & Mainen, Z. F. Optogenetic activation of dorsal raphe serotonin neurons rapidly inhibits spontaneous but not odor-evoked activity in olfactory cortex. J. Neurosci. 36, 7–18 (2016).

Scientifc Reports | (2021) 11:1680 | https://doi.org/10.1038/s41598-021-81327-z 8 Vol:.(1234567890) www.nature.com/scientificreports/

60. Siemann, J. K. et al. An autism-associated serotonin transporter variant disrupts multisensory processing. Transl. Psychiatry 7, 2 (2017). 61. Pohlkamp, T. et al. Characterization and distribution of reelin-positive interneuron subtypes in the rat barrel cortex. Cereb. Cortex 24, 3046–3058 (2014). 62. Guy, J. & Staiger, J. F. Te functioning of a cortex without layers. Front. Neuroanat. 11, 1–13 (2017). 63. Burke, M., Schwindt, W., Ludwig, U., Hennig, J. & Hoehn, M. Facilitation of electric forepaw stimulation-induced somatosensory activation in rats by additional acoustic stimulation: An fMRI investigation. Magn. Reson. Med. 44, 317–321 (2000). 64. Burr, D. & Gori, M. Te Neural Bases of Multisensory Processes. Te Neural Bases of Multisensory Processes (CRC Press, Baco Ration, 2012). 65. Dekker, T. M. et al. Late development of cue integration is linked to sensory fusion in cortex report late development of cue integration is linked to sensory fusion in cortex. Curr. Biol. 25, 2856–2861 (2015). 66. Dionne-dostie, E., Paquette, N., Lassonde, M. & Gallagher, A. Multisensory integration and child neurodevelopment. Brain. Sci. https://doi.org/10.3390/brainsci5010032 (2015). 67. Gohil, K., Bluschke, A., Roessner, V., Stock, A. & Beste, C. Sensory processes modulate diferences in multi-component behavior and cognitive control between childhood and adulthood. Hum. Brain. Mapp. 4945, 4933–4945 (2017). 68. Sieben, K., Bieler, M., Röder, B. & Hanganu-opatz, I. L. Neonatal restriction of tactile inputs leads to long-lasting impairments of cross-modal processing. PLoS Biol. https://doi.org/10.1371/journal.pbio.1002304 (2015). 69. Spenger, C. et al. Functional MRI at 47 Tesla of the rat brain during electric stimulation of forepaw, hindpaw, or tail in single- and multislice experiments. Exp. Neurol. 166, 246–253 (2000). 70. Paasonen, J., Stenroos, P., Salo, R. A., Kiviniemi, V. & Gröhn, O. Functional connectivity under six anesthesia protocols and the awake condition in rat brain. Neuroimage 172, 9–20 (2018). Acknowledgements Te support of Michael Diedenhofen with the image preprocessing is gratefully acknowledged. We tank Guil- len Fernandez for the valuable feedback on the manuscript. JH and DS were supported by a Brain & Cognition from Te Dutch research council (Grant No. 433-09-311). Funding organizations had no further role in the writing process. Author contributions N.K., D.S. and J.H. designed the study, N.K. conducted the study under supervision of D.W. and M.H., N.K. analysed the data under supervision of D.W. and M.H., N.K. wrote the manuscript under supervision of D.S. and J.H., all authors edited and approved the manuscript.

Competing interests Te authors declare no competing interests. Additional information Correspondence and requests for materials should be addressed to J.R.H. Reprints and permissions information is available at www.nature.com/reprints. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afliations. Open Access Tis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. Te images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Scientifc Reports | (2021) 11:1680 | https://doi.org/10.1038/s41598-021-81327-z 9 Vol.:(0123456789)